This paper explores the existing gaps in the way centrally sponsored schemes (CSS) are evaluated for their impact in the Indian public sector. The authors identify an ideal type model for a public sector impact evaluation mechanism that would be both effective and feasible to implement. To illustrate how incremental changes could help the existing evaluation efforts to move closer to this ‘ideal type, they undertake a thought experiment with the Pradhan Mantri Gram Sadak Yojana (PMGSY) scheme.The authors conclude with a brief forecast of why such steps might be resisted, and provide recommendations of how to overcome resistance to enable near-term improvements to the current evaluation culture.